diff --git a/tests/v1/executor/test_multiproc_executor.py b/tests/v1/executor/test_multiproc_executor.py index 4ebaba5f11..706bfae381 100644 --- a/tests/v1/executor/test_multiproc_executor.py +++ b/tests/v1/executor/test_multiproc_executor.py @@ -2,12 +2,14 @@ # SPDX-FileCopyrightText: Copyright contributors to the vLLM project import weakref +from collections import deque from types import SimpleNamespace from typing import Any import pytest -from vllm.v1.executor.multiproc_executor import WorkerProc +from vllm.v1.executor.multiproc_executor import MultiprocExecutor, WorkerProc +from vllm.v1.outputs import DraftTokenIds class _ExitWorkerLoop(RuntimeError): @@ -65,3 +67,29 @@ def test_execute_worker_rpc_returns_worker_exception(): assert len(outputs) == 1 assert isinstance(outputs[0], RuntimeError) assert str(outputs[0]) == "test error" + + +@pytest.mark.parametrize("stalled", [False, True]) +def test_take_draft_token_ids_uses_execute_model_timeout(monkeypatch, stalled): + """A wedged draft-token readback must not block EngineCore forever.""" + monkeypatch.setenv("VLLM_EXECUTE_MODEL_TIMEOUT_SECONDS", "7") + draft = DraftTokenIds(["req"], [[1, 2, 3]]) + + def dequeue(*, timeout=None): + assert timeout is not None and 0 < timeout <= 7 + if stalled: + raise TimeoutError + return WorkerProc.ResponseStatus.SUCCESS, draft + + executor: Any = MultiprocExecutor.__new__(MultiprocExecutor) + executor.is_failed = False + executor.output_rank = 1 + executor.futures_queue = deque() + executor.rpc_broadcast_mq = SimpleNamespace(enqueue=lambda payload: None) + executor.response_mqs = [None, SimpleNamespace(dequeue=dequeue)] + + if stalled: + with pytest.raises(TimeoutError, match="take_draft_token_ids timed out"): + executor.take_draft_token_ids() + else: + assert executor.take_draft_token_ids() is draft diff --git a/vllm/v1/executor/multiproc_executor.py b/vllm/v1/executor/multiproc_executor.py index 5c9f596d0c..78290d25ba 100644 --- a/vllm/v1/executor/multiproc_executor.py +++ b/vllm/v1/executor/multiproc_executor.py @@ -369,7 +369,9 @@ class MultiprocExecutor(Executor): def take_draft_token_ids(self) -> DraftTokenIds | None: # OPTIMIZATION: Get output only from a single worker (output_rank) return self.collective_rpc( - "take_draft_token_ids", unique_reply_rank=self.output_rank + "take_draft_token_ids", + unique_reply_rank=self.output_rank, + timeout=envs.VLLM_EXECUTE_MODEL_TIMEOUT_SECONDS, ) def collective_rpc( # type: ignore[override]